Elon Musk has done it again. His latest creation in the field of artificial intelligence, Grok-2, is causing a considerable stir on the web.
And no wonder: with the ability to generate images with few limits, this chatbot is leaving its mark on X, the platform formerly known as Twitter.
Here we tell you everything about the release of Grok-2, its characteristics, and the controversies that are already marking its short existence.
The launch of Grok-2: Musk’s new gem
On August 13, 2024, xAI, the artificial intelligence company owned by Elon Musk, launched Grok-2 and its mini version in beta. Grok-2 is billed as a significant evolution of its predecessor, Grok-1.5, and promises advanced capabilities in chat, coding, and reasoning.
But what has really captured users’ attention is the inclusion of a functionality to generate images directly in X, accessible only to Premium and Premium+ subscribers.
According to xAI, Grok-2 is a big step forward, offering a preview of what artificial intelligence can do in terms of interaction and visual creativity.
Since its launch, users have taken advantage of this feature to create and share images ranging from the surreal to the controversial, without significant moderation involved.
Grok-2 Features: More than just a chatbot
Grok-2 is distinguished by its ability to generate images from text, a function that has been enabled thanks to the collaboration with Black Forest Labs, a German startup that was only presented “in society” on August 1, 2024.
This young company has developed the FLUX.1 model, which powers the imaging function of Grok-2. According to some comparisons, the images generated by FLUX.1 surpass in quality those of other leading models such as Midjourney and OpenAI.
However, the most notable—and worrying—feature of Grok-2 is the lack of safeguards in the generation of these images.
While other industry majors such as Google and OpenAI have implemented restrictions to prevent the creation of inappropriate or dangerous content, Grok-2 appears to operate without many of these limitations.
This has allowed users to generate images of famous people in compromising situations, and even violent or potentially illegal scenes, which has unleashed a wave of criticism.
Controversies around Grok-2: Freedom of expression or irresponsibility?
Since its launch, Grok-2 has been at the center of several controversies. One of the first alarms was raised by the chatbot’s ability to generate images of political figures without any type of restriction.
With the US presidential election around the corner, many fear that Grok-2 will become a tool to spread visual disinformation.
Furthermore, the very nature of the images generated has raised eyebrows. Examples such as images of Donald Trump smoking marijuana on the Joe Rogan show, or Pikachu carrying an assault rifle, have gone viral on X.
These creations, along with others involving copyrighted characters, suggest that the FLUX.1 model has been trained without respecting the typical constraints seen in other AI models.
But not everything is fantasy or surrealism. Grok-2 has been criticized for the ease with which images can be generated that border on violence or inappropriateness.
Media such as The Guardian have shown how it is possible to create images of public figures such as Kamala Harris or Taylor Swift in underwear, something that has undoubtedly generated a debate about the ethical limits of this technology.
Grok-2 mini: Grok-2’s little brother
Along with the launch of Grok-2, xAI introduces Grok-2 mini, a lighter but more powerful version. This model, although more compact compared to its older brother, stands out for its notable balance between execution speed and quality of responses.
Grok-2 mini has been designed to handle tasks efficiently, while maintaining high precision in its results, making it an ideal choice for those looking for performance without compromising quality.
This tool also inherits the ability to generate images from text prompts, a feature that has already caused a stir on social network X due to its flexibility and potential for creativity.
Like Grok-2, this smaller version is available in beta for X Premium and Premium+ subscribers, allowing you to explore its capabilities early.
The presence of Grok-2 mini underlines xAI’s intention to offer AI solutions adapted to different needs, from users who require high processing speed to those looking for a versatile and accessible tool for various applications.
Criticism and reactions to the “funniest AI” according to Musk
The criticisms have not been long in coming. Alejandra Caraballo, an instructor at the Harvard Cyber Law Clinic, called Grok-2 “one of the most irresponsible AI implementations I have ever seen.”
This opinion reflects the sentiment of many in the technology community, who see Grok-2 as a dangerous precedent in the race for AI supremacy.
As if that were not enough, X and Musk are facing several legal fronts. Nine privacy complaints have been filed against the platform in the European Union, accusing X of using user data to train Grok-2 without their consent.
These complaints, backed by privacy rights organization NOYB, highlight possible violation of the General Data Protection Regulation (GDPR) and could result in significant fines for the company.
Musk himself has responded to the criticism in his characteristic style, emphasizing that Grok is “the funniest AI in the world.” But for many, the fun comes at too high a price, especially when it comes to misinformation and user privacy.
Can I try Grok from Spain?
Yes, you can try Grok from Spain. Grok has been available to users in the European Union since May 2024. Initially, it was only available in the United Kingdom, but has now expanded to other EU countries. However, you need to be an X Premium subscriber to access Grok.
Does Grok represent the future of AI?
Without a doubt, the new Grok represents a milestone in the evolution of chatbots and AI imaging, but it is also a reminder of the ethical challenges that accompany these technological advances.
With Musk betting on a philosophy of “freedom of expression” without many restrictions, the future of Grok-2 is full of unknowns.
Will Grok-2 be a tool to boost creativity on social media? Or will it become a focus of misinformation and legal controversies? The only certainty is that Grok-2 has already left its mark, and we will probably continue to hear it in the coming months.
In this sense, Grok-2 mini also plays a crucial role, offering a more agile and accessible alternative for those looking to experiment with AI capabilities without sacrificing the quality of their results.
Therefore, X users will continue to explore the limits of this new tool, with all the implications that may have.
This post is also available in: